I've just spent about an hour on the survey, at which point I noticed the progress bar was at about 1/6th. This was at the start of four timed questions which required 2 minutes each, with each one having a few follow-up questions.

At this point there had been 7 or 8 of these timed questions, as well as at least two dozen pages of multiple-choice questions and a few 'brain-teasers'. I do not see how it is possible to complete this first 1/6th in under 45 minutes, seeing that the timed questions alone already take up 16 minutes.

This post seems to confuse Effective Altruism, which is a methodology, for a value system. Valuing the 'impartial good' or ' general good' is entirely independent of wanting to do 'good' effectively, whatever you may find to be good.

You articulate this confusion most clearly in the paragraph starting "Maybe it would help to make the implications more explicit." You make two comparisons of goals that one can choose between (shrimp or human, 10% chance of a millions lives, or 1000 lives for sure). But the value of the options is not dictated by effective altruism; this depends on ones valuation of shrimp vs human life in the first case, and ones risk profile in the second.